Jump to content
UltravioletPhotography

[Filter Test] Baader UV/IR-Cut, B+W 38/39/470, Schott BG40/S8612: IR-Blockers


Andrea B.

Recommended Posts

EDITOR'S NOTE: 10 Dec 2021

I currently favor Schott BG38 x 2.0 mm for IR-blocking.

LINK:  https://www.ultravioletphotography.com/content/index.php?/topic/4776-five-ir-blockers-tested-for-visible-color-use-on-full-spectrum-camera/page/4/#comment-50176


 

 

You can skip the formal write-up of the test here in the first post. The photos begin in the next post.

Here's the conclusion.

 

BOTTOM LINE:

1. Combine an in-camera preset white balance with a camera+lens+filter colour profile for the most accurate colour in Visible photographs made with a converted ('full spectrum' = 'broadband') digital camera.

2. The Baader UVIR-Cut and the BG40 produced the most accurate colours when colour profiles were combined with a pre-set white balance for the D600-BB + CO60/4.0 + Filter combo. Keep in mind that a blue-green filter like the BG40 transmits UV. So in spite of the BG40 colour accuracy in these tests, it may give a hazy or foggy look in some shooting scenarios.

 

EDIT: 13 Dec 2017.

A hazy/foggy appearance often called "UV-haze" can occur in photos made at high altitudes containing long distances due to scattered UV/violet/blue light. UV-haze can sometimes occur in seashore scenes made out across the ocean. UV-haze is due to scattered UV/violet/blue light and not due to particulate matter in the air. In film days a "UV-filter", which we would now call a UV-blocker, was used to block some of the unwanted scattered light which softened detail in the photograph. Now our digital cameras contain internal filtration which blocks UV-haze so some digital photographers may never have experienced the phenomenon. Anyone using a converted camera for visible work may experience UV-haze if their UV-blocker is not strong enough. Given that BG glass does not block UV, a tighter filtration might be needed for conversions when used for high-altitude scenarios.

 

 


 

 

Test: UV/IR-Block & IR-Block Filters on a Broadband Converted Digital Camera

 

Background: During conversion, the internal UV/IR-Block filters are removed from the camera. To make a Visible light photo with a converted camera, an external UV/IR-Block or IR-Block filter must used on the lens. Unless the external blocker is identical to the one removed, the camera's Visible colours may be altered. Let's look at that.

 

Reference:

NIR cutoff filter for true color imaging sensors

by Ralf Biertümpfel and Steffen Reichel

[both authors worked for Schott AG Advanced Optics]

Advanced Optical Technologies

Editor-in-Chief: Pfeffer, Michael

Volume 2, Issue 5-6 (Dec 2013)

ABSTRACT

The function of a near-infrared (NIR) cutoff filter for imaging sensors is being described.

The main purpose of the NIR cut filter is to obtain correct color recognition;

therefore, the NIR filter is made of an absorbing filter glass and an interference coating.

The absorbing filter glass is needed to minimize multiple reflections inside the lens system,

which are the cause for ghost images. An additional interference coating enhances

the function of the filter. Coating and filter glass are strongly dependent on each other.

This requires high reproducibility and low tolerances of the filter glass and interference coating.

In addition, features like inner quality – especially striae – and stability

of the refractive index are important. A NIR cut filter may be designed as a flat plate

or as a lens. Our analysis provides an estimation about striae level and variation of

transmittance and their effect on image quality and color recognition. Furthermore, the use

of an absorption filter glass as a lens (shrinking down the overall size) is discussed in

terms of the influence on transmission and striae.

 

UV/IR-Block and IR-Block Filters:

Here are the six filters which I have.

  • Baader UV/IR-Cut
  • B+W 38 = IR-Block, BG38 (I'll use the BG 38 designation in what follows.)
  • B+W 39 = IR-Block, BG39 (I'll use the BG 39 designation in what follows.)
  • Schott BG40, IR-Block
  • B+W 470, IR-Block (I do not know anything about this filter. Call it BG470 below.)
  • Schott S8612, IR-Block

Camera + Lens:

The Coastal Optics 60/4.0 is UV/Vis/IR apochromatic and very colour accurate.

  • Nikon D600-Broadband + Coastal Optics 60/4.0

Standards:

  • Color Checker Passport for color profiling.
  • Labsphere 99% and %75 reflective Spectralon standards for white balance.

Procedure 1:

Make a visible photo of the CC Passport and the standards using the camera's auto white balance setting for each filter.

 

Convert photos in Capture NX2 to preserve the camera's auto white balance settings as shot.

Save the converted As Shot photos.

 

Then in NX2 re-set the white balance over the 99% white reflective standard.

Save the Re-balanced photos.

 

Then, out of sheer curosity, convert the photos in Raw Digger to see what colour were actually recorded.

Save the converted Raw Composite photos.

 

For each filter compare the As Shot auto white balanced photo, the Re-balanced photo and the Raw Composite.

 

How well did auto white balance perform as compared to the re-balanced photo?

We would expect there to be some large differences here because auto WB under a color filter may not be able to fully compensate.

 

Which filter, if any, performed best under auto WB?

 

(And what colours were actually recorded?)

 

 

Procedure 2: Pre-set an in-camera white balance for each filter against the white reflective standard.

Make a visible photo of the CC Passport and the standards using this pre-set white balance for each filter.

 

Convert photos in Capture NX2 to preserve the camera's auto white balance settings as shot.

Save the converted As Shot photos.

 

Then in NX2 re-set the white balance over the 99% white reflective standard.

Save the Re-balanced photos.

 

Then, convert the photos in Raw Digger to see what colour were actually recorded.

Save the converted Raw Composite photos.

 

For each filter compare the As Shot pre-set white balanced photo, the Re-balanced photo and the Raw Composite.

 

How well did the in-camera Pre-set white balance perform as compared to the Re-balanced photo?

We would expect the results to be almost identical.

 

Which filter performed best under pre-set WB?

We would expect all photos have the same WB and thus there would be no 'best'.

 

(And what colours were actually recorded?)

 

Procedure 3:

Make and apply a color profile in Photo Ninja for each filter using the raw NEFs from Procedure 2.

 

Attempt to set the saturation level in the profiled version of each photo to match the saturation level preserved in the NX2 conversion of the Procedure 2 photos

 

Compare the profiled photos to the pre-set WB photos.

 

Are there changes in the colours when colour profiling is added to a pre-set white balance?

Which filter, if any, performs best under pre-set WB combined with colour profiling?

 

Exposure Settings:

  • f/8 @ ISO-100 with varying shutter speeds, as needed.
  • Metering: Matrix.
  • Mode: Manual.
  • Nikon Picture Control: Neutral with all settings = 0 except for Sharpening = 4.
  • White Balance: Auto or Pre-set for each individual filter as described above.

Exposure Details:

To obtain similar exposures for each tested filter, I initially shot at the metered shutter speed. I then slowed the shutter speed in 1/3 increments to push the luminosity histogram to the right until I began to get white blinkies. (Yes, we all know this is a JPG luminosity histogram and not a true raw histogram, but I've had a lot of practice pushing exposures this way without blowing out the 99% white standard.)

I observed that the metered shutter speed tended to underexpose through the blue-green filters. This is not a surprise.

 

Conversion/Edit Details:

The raw files were converted in Capture NX2 to preserve the white balance settings. The sRGB assignment was made. The Neutral Picture Control was confirmed. No other edits were made. Files were saved as TIFs.

 

The NX2 conversion was then repeated as described with an additional step: a marquee white balance was made on the 99% reflective white standard. Files were again saved as TIFs.

 

A final conversion was made in Raw Digger to extract the raw composite of the file with no white balance applied. Files were again saved as TIFs.

 

All TIFs were rotated and cropped in Photo Ninja.

All TIFs were then resized and saved as JPGs in Photo Mechanic.

The composite JPG strips were created in Photoshop Elements.

Link to comment

Procedure 1: Auto White Balance for Each Filter

 

Row 1: As shot with camera's auto white balance.

Row 2: New white balance applied in NX2 on the white standard.

Row 3: Raw composite extracted in Raw Digger with no white balance applied.

 

Filters From Left to Right: Baader UV/IR-Cut, BG38, BG40, BG39, S8612, BG470

The grouping goes from pink on the left to cyan on the right.

 

Row 1 :: Best Filter with Auto White Balance = BG40.

Remember, this applies only to the Nikon D600 + CO60/4.0.

You might get different results with another camera + lens combination.

The BG40 auto white balance photo is very, very close to its re-balanced photo (just below it).

The BG38 was fairly good, but shows a pale pinkish tint with auto WB.

The Baader UVIR-Cut shows a pronounced pink tint with auto WB.

The BG39, S8612 and BG470 all show a strong cyan tint with auto WB.

 

Row 3 :: Raw Composite Observations

I took some samples on the 99% reflective white standard and found that the Baader UV/IR-Cut raw composite (on the bottom left) shows a yellow-green tint. The BG38 and BG40 both show a green-blue tint (more G than blue). The remaining filters all show a cyan tint.

 

Row 2 :: Color Differences in the Re-balanced Photos of the Middle Row

This was very interesting. At first glance, all 6 filters seem to produce the same colours when white balance is re-set in the NX2 converter. But if you look closely, you can see that there are colour differences between all 6 photos in the middle row.

Look in particular at the red patch on the CC Passport -- row 3, column 3. As your eye travels from left to right over that red patch, you can see that the red changes from an orange-red to a red to a dark, almost brownish red. The purple patch in row2, column 4 also shows much variation as you go from left to right.

 

Most Accurate IR-Block Filter for D600-BB after NX2 White Balance: BG40 - with some reservations - so I'll still be using a UV/IR-Cut.

 

I don't think we can answer the 'most accurate' question in its entirety because we must take into account all the variables which can affect our colour perception: ambient light, monitor brightness and accuracy, the sRGB/JPG shifts and so forth.

 

However, an informal comparison made by holding the CC Passport next to the 2nd row photos on the monitor indicates that the BG38 (good) and the BG40 (better) seem to provide the most accurate colours after white balance is adjusted in NX2 during conversion. Remember that this conclusion is applicable only to the D600-broadband.

 

And let's not forget the Spanner in the Works -- that BG40 and its other blue-green friends transmit both Visible and UV light. If we are using a BG filter as a colour correction filter on a converted camera, are we going to encounter situations where the Vis and the UV do not play well together? Well, yes we are. Rather often I think, because we will be seeing quite a lot of UV haze when using a blue-green filter for certain subjects. Medium to long distant landscapes come to mind, for one such scenario.

 

This composite photo is 948 x 600 pixels when clicked in an expanded browser.

autoWB_3strip.jpg

Link to comment

Procedure 2: Pre-set White Balance for Each Filter

 

Row 1: As shot with pre-set white balance made on the 99% white standard.

Row 2: New white balance applied in NX2 on the 99% white standard.

Row 3: Raw composite extracted in Raw Digger with no white balance applied.

 

Filters From Left to Right: Baader UV/IR-Cut, BG38, BG40, BG39, S8612, BG470

 

Row 1 and Row2 :: Each filter photo in the top row should look the same as the corresponding filter photo in the second row because a Nikon in-camera pre-set white balance can be preserved in the Nikon converter (NX2) and re-setting that white balance in NX2 would only result in the most minor of changes due to sampling location error.

 

Row 3 :: These raw composite shots would be the same as those made above in Procedure 1 (subject to very minor exposure differences). Thus the remarks above about the various color casts still pertain.

 

Colour Differences between the Filters after White Balance

Again, a casual glance could cause you to think the colours produced under each filter were the same. And several colours do seem almost the same in Rows 1 and 2: the yellows, greens and oranges, for example. But look at the reds, magentas and blues. They vary quite a lot.

 

Most Accurate IR-Block Filter for D600-BB Pre-set White Balance: BG40 - with some reservations - so it is still an UV/IR-Cut for me.

 

As before, I don't think we can answer the 'most accurate' question in its entirety because we must take into account all the variables which can affect our colour perception: ambient light, monitor brightness and accuracy, the sRGB/JPG shifts and so forth.

 

Again, an informal comparison made by holding the CC Passport next to the 2nd row photos on the monitor indicates that the BG38 (good) and the BG40 (better) seem to provide the most accurate colours when using an in-camera pre-set white balance. Remember that this conclusion is applicable only to the D600-broadband.

 

But we still have the same Spanner as before. UV haze might interfere with good clear photo making when using a blue-green BG40 filter for colour correction.

 

presetWB_strip.jpg

Link to comment

Procedure 3: Profiled Colour + Pre-set White Balance

 

I got a bit tired of all this conversion and editing, so I decided to make colour profiles for only 3 of the 6 filters used in the previous two experiments. I picked the often used Baader UVIR-Cut, the accurate BG40 and one filter from the too-much-cyan group, namely the S8612.

 

Row 1: As shot with pre-set white balance made on the 99% white standard.

Row 2: Added colour profile made for camera + lens + filter combination.

 

Filters From Left to Right: Baader UV/IR-Cut, BG40, S8612

 

Conclusion 1:

The 3 WB-with-Colour-Profile photos in the 2nd row look more alike and more colour-accurate than the WB-only photos in the 1st row which show variation in colours between the 3 photos (especially the reds, blues, magentas and purples). So colour profiling is an important part of making good Visible photos with a converted camera.

 

Conclusion 2:

Of the 3 WB-with-Colour-Profile photos in the 2nd row, the UVIR-Cut filter and the BG40 filter are the most colour accurate. The S8612 does not quite fully recover with the addition of a colour profile to WB, although it is very very close to the other two.

 

[i would expect the BG38 to look most like the BG40. And the BG470 and BG39 to turn out like the S8612. I will try to add them later when I have some time.]

 

presetWBprofile_strip.jpg

Link to comment

Agree with Steve (Cadmium). Outstanding work and paying all the necessary attention to details too.

 

I have had excellent visible-light captures with the Baader UV/IR Cut filter after making a dedicated profile for it in Photo Ninja. Without profile it will not be accurate in colour rendition and makes a reddish stamp on everything. Thus confirming Andrea's findings.

 

BG-38 didn't work well, so I dropped using it for the purpose of making VIS references to UV captures.

 

The above applies when I'm using a broad-band converted camera for UV and visible-light work. I often find it better to use two separate cameras of the same sensor format (m43, DX, or FX) combined with two lenses matching in focal length. However, for real close-ups, the one camera approach is better as I then always use a tripod-mounted set up. Swapping filters is quick and easy if a filter box or bayonet-mounted filters are used.

Link to comment

Thank you everyone.

 

It was an interesting experiment. And I may further investigate that BG40.

 

Somewhere in the pages of various Schott technical publications or catalogs or maybe online, there is a statement about using BG glass for colour balance with digital sensors. And we have seen the BG filters that come out of some digital cameras. What kind of BG, however, we do not know. Or whether it is coated. Or how it is combined with UV blocking.

 

Thus the topic of restoring proper Visible colour to a full spectrum camera still has some unanswered questions. Of course, colour profiling and proper white balance get us close even though there is still some variability with respect to saturation, contrast and viewing parameters.

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...